EP0604295B1 - A method for automatic gain control of a digital receiver, in particular in a TDMA receiver and a device for performing the method - Google Patents

Abstract

Method of automatic gain control of a digital receiver (1) comprising amplifier means (2). The method comprises the following steps: - sampling at a predetermined sampling frequency of the output signal (s(t)) from the said amplifier means; - processing of the samples obtained including, for each sample, a comparison of the absolute value of this sample respectively to a predetermined maximum and minimum threshold; and - adjustment of the gain (G) of the amplifier means (2) carried out on completion of each sample processing, in such a way that: * if a sample is outside a maximum bracket [-SMAX, +SMAX], this gain (G) is reduced by a predetermined reduction level, * if several successive samples, of a predetermined number, lie within a minimum bracket [-SMIN, +SMIN], this gain (G) is increased by a predetermined increase level. Use especially in mobile telephony. <IMAGE>

Description

The present invention relates to an automatic gain control method for a digital receiver, in particular a time division multiple access (TDMA) receiver.

The digital frames received are composed of data packets transmitted in turn by a set of more or less distant transmitters. However, two successive packets may have a significant level difference on reception, for example 90 dB. The gain of the reception amplifier must be adjusted to bring the signal level into the dynamics of the processing chain that follows the amplifier.

• une rampe de montée en puissance avec possibilité d'une légère sur-oscillation, d'une durée de 28 µs ;
• une période durant laquelle la puissance est constante, d'une durée de 542,8 µs durant laquelle l'information est transmise. Cette période sera dénommée "période utile" par la suite; et
• une rampe de baisse de puissance d'une durée de 28 µs.

A data packet includes as an example:

• a ramp up with the possibility of a slight over-oscillation, lasting 28 µs;
• a period during which the power is constant, of a duration of 542.8 µs during which the information is transmitted. This period will be called "useful period"hereafter; and
• a 28 µs duration power drop ramp.

The ramp down of one data packet may partially overlap the ramp up of the next packet. However, the gain of the reception chain must be adjusted at the arrival of each data packet in order to reduce the signal level during the useful period in the dynamics of the processing chain located downstream of the receiver. This gain must remain constant during the reception of the useful period in order to optimize the operation of the baseband processing.

Up to now, to overcome this problem, a first solution consisted in having several reception amplifiers presenting different gains and working in parallel. The output of each of these amplifiers was sampled and then stored, and a selection member came to choose the best stored result. Another solution consisted in using a delay line followed by an amplifier controlled by a device taking the signal received upstream of the delay line, analyzing the power of this signal and adjusting the gain of the amplifier in order to properly amplify this same signal at the output of the delay line.

However, the selection technique which involves the use of several amplifiers is an expensive solution. Furthermore, the use of a delay line followed by an amplifier leads to noise generation and a delay.

The object of the present invention is to remedy these drawbacks by proposing an automatic gain control method which does not induce noise or delay while being less expensive than existing methods.

• un échantillonnage à une fréquence d'échantillonnage prédéterminée du signal de sortie desdits moyens amplificateurs ;
• un traitement des échantillons obtenus comportant, pour chaque échantillon, une comparaison de la valeur absolue de cet échantillon respectivement à un seuil maximum et à un seuil minimum prédéterminés associés respectivement à une fourchette maximum [-SMAX, +SMAX] et une fourchette minimum [-SMIN, +SMIN] et
• un réglage du gain des moyens amplificateurs effectué à l'issue de chaque traitement d'échantillon, de sorte que:
  • * si un échantillon est en dehors de la fourchette maximum [-SMAX, +SMAX], ce gain est réduit d'un niveau de réduction prédéterminé ;
  • * si plusieurs échantillons successifs en nombre prédéterminé sont compris dans la fourchette minimum [-SMIN, +SMIN], ce gain est augmenté d'un niveau d'augmentation prédéterminé.

According to the invention, the automatic gain control method for a digital receiver, in particular a digital time-division multiple access receiver (TDMA), this receiver comprising adjustable gain amplifier means receiving an input signal, is characterized in that it comprises the following stages:

• sampling at a predetermined sampling frequency of the output signal of said amplifier means;
• a processing of the samples obtained comprising, for each sample, a comparison of the absolute value of this sample respectively to a predetermined maximum threshold and to a minimum threshold associated respectively with a maximum range [-SMAX, + SMAX] and a minimum range [- SMIN, + SMIN] and
• an adjustment of the gain of the amplifying means carried out at the end of each sample processing, so that:
  • * if a sample is outside the maximum range [-SMAX, + SMAX], this gain is reduced by a predetermined reduction level;
  • * if several successive samples in predetermined number are included in the minimum range [-SMIN, + SMIN], this gain is increased by a predetermined increase level.

According to a first embodiment of the method according to the invention, each sample is compared with four thresholds: a maximum threshold and a minimum threshold and their respective opposite values.

According to a second embodiment of the method according to the invention, the absolute value of each sample is determined and then compared to the maximum threshold and the minimum threshold.

According to another aspect of the invention, the automatic gain control device for a digital receiver, in particular a digital time-division multiple access receiver (TDMA), this receiver comprising adjustable gain amplifier means receiving a signal d input, implementing the method according to the invention, is characterized in that it comprises means for sampling at a predetermined sampling frequency the output signal of said amplifying means, means for comparing the absolute value of each sample obtained respectively at a maximum threshold and at a minimum threshold and for generating threshold detection information, and means for adjusting the gain of the amplifying means cooperating with the comparison means to reduce said gain as soon as a predetermined reduction level the sample is detected outside the maximum range [-SMAX, + SMAX], and to increase a predetermined increase level of said gain as soon as several samples in predetermined number are included in the minimum range [-SMIN, + SMIN].

Thus, the present invention has the advantage of allowing integration of the automatic gain control device in an integrated circuit and of using a very simple detector for the feedback loop. In addition, the stability of the system is increased by the control logic implemented with the method according to the invention. In addition, a correction of 90 dB is obtained in the GSM for a bit period and the method according to the invention operates in voltage and not in power, which contributes to improving the speed of the automatic gain control.

• la figure 1 est un schéma synoptique illustrant un dispositif de contrôle automatique de gain selon l'invention ;
• la figure 2A est un chronogramme correspondant au cas où le niveau initial du signal est supérieur au seuil maximum ;
• la figure 2B est un chronogramme correspondant au cas où le niveau initial du signal est inférieur à un seuil minimum ; et
• la figure 3 est un tableau donnant les probabilités respectives de dépassement d'un échantillon et d'au moins un dépassement parmi deux échantillons successifs nécessaires pour le réglage de l'algorithme et de la fréquence d'échantillonnage mettant en oeuvre le procédé selon l'invention dans le cas d'un signal d'entrée sinusoïdal d'amplitude A.

Other features and advantages of the invention will appear in the description below. In the appended drawings given by way of nonlimiting examples:

• Figure 1 is a block diagram illustrating an automatic gain control device according to the invention;
• FIG. 2A is a timing diagram corresponding to the case where the initial level of the signal is greater than the maximum threshold;
• FIG. 2B is a timing diagram corresponding to the case where the initial level of the signal is below a minimum threshold; and
• FIG. 3 is a table giving the respective probabilities of overshoot of a sample and of at least one overshoot among two successive samples necessary for the adjustment of the algorithm and of the sampling frequency implementing the method according to invention in the case of a sinusoidal input signal of amplitude A.

We will now describe an example of implementation of the method according to the invention at the same time as the associated device.

Within an automatic gain control device 1 according to the invention, an amplifier 2 which works at intermediate frequency (IF) receives an analog signal e (t) in the form of a succession of data packets and is monitored at the output of a signal amplitude detector 4 with two thresholds (minimum and maximum) at the input of which is placed a sampler 5. An analog / digital converter 6 is also placed at the output of amplifier 2 to deliver a set of digitized data N which are subject to downstream processing. The samples collected at the output of the detector 4 feed a logic control device 3 which regulates the gain of the amplifier 2 so as to prevent any risk of oscillation.

• si la valeur absolue d'un échantillon dépasse un seuil maximum SMAX, à savoir si cet échantillon est en dehors de la fourchette [-SMAX, +SMAX], le gain est diminué de 3 dB ;
• si deux échantillons successifs sont compris dans la fourchette minimum [-SMIN, +SMIN], le gain est augmenté de 3 dB.

By way of example, a control algorithm embodying the method according to the invention and implemented within the logic control device 3 can comprise the following steps:

• if the absolute value of a sample exceeds a maximum SMAX threshold, ie if this sample is outside the range [-SMAX, + SMAX], the gain is reduced by 3 dB;
• if two successive samples are included in the minimum range [-SMIN, + SMIN], the gain is increased by 3 dB.

• e(t)=A(t).sin [ω₀ t+ϕ(t)] ;
  où:
     A(t) est conforme à des masques de puissance prédéterminés, notamment des masques prescrits dans une recommandation technique telle "GSM 05.05'' ;
• ω₀ est la pulsation de la fréquence intermédiaire IF ; et
• ϕ(t) est la phase modulée par le flux de données.

The signal e (t) at the input of amplifier 2 can be modeled in the following form:

• e (t) = A (t) .sin [ω ₀ t + ϕ (t)];
  or:
  A (t) complies with predetermined power masks, in particular masks prescribed in a technical recommendation such as "GSM 05.05 '';
• ω ₀ is the pulsation of the intermediate frequency IF; and
• ϕ (t) is the phase modulated by the data flow.

As an example of a practical embodiment of the invention, the variations of A (t) are slow, for example, less than 0.4 dB between two samples, and the intermediate frequency can for example be equal to 9.75 MHz.

• u(kTe)=sin[ω₀kTe+ϕ(kTe)].

The sampler 5 operates at a frequency Fe of 13 MHz, namely 48 times the symbol frequency, in the case of GSM. the output samples are of the type:

• u (kTe) = sin [ω ₀ kTe + ϕ (kTe)].

As the phase ϕ (t) varies by at most π / 2 between two symbols in the case of GSM, it will vary by less than π / 96 between two samples. The phase variation due to the carrier frequency, on the other hand, is much greater, since it is equal to Fe ₀ / Fe, ie 1.5π.

The phase variation between two successive samples is (taking samples k and k + 1 as an example): $${\text{ω}}_{\text{0}}\text{Te + ϕ ((k + 1) Te) -ϕ ((k-1) Te)}$$

• variation due à la fréquence intermédiaire: ω₀Te, c'est à dire, ω₀/Fe ;
• variation due à la modification par le flux de données:ϕ((k+1)Te)-ϕ((k-1)Te).

We observe that it breaks down into two terms:

• variation due to the intermediate frequency: ω ₀ Te, ie, ω ₀ / Fe;
• variation due to the modification by the data flow: ϕ ((k + 1) Te) -ϕ ((k-1) Te).

The sampling frequency Fe and the intermediate frequency (of pulsation ω ₀ ) are chosen so that the ratio ω ₀ / Fe is an odd multiple of π / 2.

The sampling frequency Fe is also chosen to be much higher than the symbol frequency of the signal received at the input so that the phase variation due to the modulation by the data flow between two successive samples is negligible compared to π / 2.

It is moreover desirable that the sampling frequency Fe is chosen sufficiently large so that the amplitude A (t) of the signal as defined above varies slowly between two samples.

• un seuil maximum, placé par exemple à 6 dB sous le niveau de saturation de la partie bande de base du récepteur, sert à détecter les signaux trop forts, donc les gains de l'amplificateur 2 trop élevés ;
• un seuil minimum SMIN, placé par exemple à 10 dB sous le seuil maximum, sert à déterminer si le signal a une amplitude insuffisante, donc si le gain de l'amplificateur 2 est trop faible.

The threshold detector 4 evaluates the absolute value of the amplitude of each sample with respect to two thresholds:

• a maximum threshold, placed for example at 6 dB below the saturation level of the baseband part of the receiver, is used to detect signals that are too strong, therefore the gains of amplifier 2 too high;
• a minimum threshold SMIN, placed for example at 10 dB below the maximum threshold, is used to determine if the signal has an insufficient amplitude, therefore if the gain of amplifier 2 is too low.

• si la valeur absolue d'un échantillon dépasse le seuil maximum, le gain est réduit de 3 dB ;
• si la valeur absolue de deux échantillons successifs est inférieure au seuil minimum, le gain est augmenté de 3 dB ;
• sinon, le gain est maintenu inchangé.

The logic control device 3 interprets the information coming from the detector 4 to adjust the gain of amplifier 2 while ensuring the stability of the loop, according to the following algorithm:

• if the absolute value of a sample exceeds the maximum threshold, the gain is reduced by 3 dB;
• if the absolute value of two successive samples is less than the minimum threshold, the gain is increased by 3 dB;
• otherwise, the gain is kept unchanged.

We will now describe an example of simulation of the above algorithm in the case of a sinusoidal input signal: e (t) = A.sin (ω ₀ t + ϕ), taking into account the different possible detection situations .

If, with reference to FIG. 2A, the initial level of the signal is greater than the maximum threshold SMAX, the probability of a threshold being exceeded as a function of the ratio between the peak voltage A and the maximum threshold is given by the table in the figure 3. The probability values indicated in this table are optimal of the ratio between the carrier frequency and the sampling frequency which allows the phase to rotate by an angle π / 2 (modulo π) between two samples.

Therefore, at a value of | sin [(ω ₀ kT)] | close to 0, corresponds to a value of | sin [(ω ₀ (k + 1) T)] | close to 1, explaining the very high probabilities in the last column of the table. It should be noted that a high value of the sampling frequency has the effect of making the phase shift between two samples negligible compared to the phase shift of π / 2 due to the carrier. In addition, variations in the amplitude A (t) between two samples are also negligible.

In the example illustrated in FIG. 2A, corresponding to a carrier frequency of 9.75 MHz and a sampling frequency of 13 MHz, it is assumed that at the initial time t = 0, the gain G of the amplifier 2 is such that the amplitude of the output signal is greater than SMAX and that the first sample taken E0 is almost zero, for non-limiting example. At time t = Te, the sample E1 is between the thresholds SMIN and SMAX and the gain of the amplifier 2 remains unchanged (gain variation ΔG = 0). At time t = 2Te, the sample E2 has a module greater than SMAX, which causes a reduction of 3 dB in the gain of amplifier 2, and consequently a reduction in the amplitude of the output signal. Sample E3 is between the minimum threshold SMIN and the maximum threshold SMAX. The gain therefore remains unchanged. The following samples E4, E5, E6, E7 are also included in absolute value in the SMIN-SMAX range. It may happen that a sample falls in the range [-SMIN, + SMIN], but this can never be the case for two successive samples. The gain G therefore remains constant for the entire duration of the packet.

• si la valeur crête du signal reçu est juste au dessus du seuil minimum SMIN, alors les valeurs absolues de deux échantillons successifs vont passer sous le seuil avec une très forte probabilité, comme c'est le cas des deux échantillons successifs F4 et F5, et le gain G de l'amplificateur 2 est alors accru de 3 dB supplémentaires ;
• si le signal est déjà bien au dessus du seuil minimum SMIN, par exemple d'au moins 3 dB, le gain de l'amplificateur 2 est maintenu constant.

If, with reference to FIG. 2B, the initial level of the signal is lower than the minimum threshold SMIN, the first sample F0 will of course be located in the range [-SMIN, + SMIN], but the automatic gain control device according to the invention will not react for reasons of stability. If the second sample F1 is also included in the range [-SMIN, + SMIN], the gain G is then increased by 3 dB for each new sample F1 under the minimum threshold SMIN. When the absolute value of a sample crosses the minimum threshold SMIN, as is the case of samples F2 and F3, the growth of the gain G will be stopped. Two cases are then possible:

• if the peak value of the received signal is just above the minimum threshold SMIN, then the absolute values of two successive samples will fall below the threshold with a very high probability, as is the case of the two successive samples F4 and F5, and the gain G of amplifier 2 is then increased by an additional 3 dB;
• if the signal is already well above the minimum threshold SMIN, for example at least 3 dB, the gain of amplifier 2 is kept constant.

In the case illustrated in FIG. 2B, a third adjustment of the gain of the amplifier 2 is still necessary to obtain a suitable stabilization of the amplitude of the output signal, since the absolute values of two successive samples F8, F9 are less than the threshold minimum SMIN.

If the initial signal level is between the two minimum thresholds SMIN and maximum SMAX, the probability of exceeding the maximum threshold is zero, and the probability of going below the minimum threshold is given by the table in FIG. 3. If the value signal peak exceeds the minimum SMIN threshold by at least 3 dB, the gain will not vary because the probability that the absolute values of two successive samples are below the threshold is zero. If the peak value is greater than the minimum threshold SMIN by less than 3 dB, it is possible that the absolute values of two successive samples are below the threshold, which causes with the method according to the invention an increase in the gain of 3 dB, driving the system into a stable area.

The simulation results presented above demonstrate that the automatic gain control method according to the invention makes it possible to recover a level variation without inducing any significant delay. It is thus possible to obtain a correction of a power jump of 90 dB over a reception time of one bit.

Of course, the invention is not limited to the examples which have just been described and numerous modifications can be made to these examples without departing from the scope of the invention.

Thus, the aforementioned minimum and maximum threshold values have been given only as a practical example, but may be chosen according to the constraints specific to the receiver concerned. In addition, the number of successive samples below the minimum threshold before an increase in amplifier gain can be greater than 2. Furthermore, the number of detection and comparison thresholds can be increased by the value of the amplifier adjustment steps, so obtain several modes, such as a catch-up mode and a chase mode. Other relationships can also be provided between the carrier frequency and the sampling frequency or between the bit frequency and the sampling frequency.

The automatic gain control method according to the invention can advantageously be implemented in a receiver for GSM and TETRA systems.

Claims (10)

1. Automatic gain control method for a digital receiver (1), especially a time division multiple access (TDMA) digital receiver, comprising variable gain amplifier means (2) receiving an input signal (e(t)), characterised in that it comprises the following steps:

   - sampling the output signal (s(t)) of said amplifier means (2) at a predetermined sampling frequency,

   - processing each sample obtained by comparing its absolute value with a predetermined maximum threshold (SMAX) and a predetermined minimum threshold (SMIN) respectively associated with a maximum range [-SMAX, +SMAX] and a minimum range [-SMIN, +SMIN], and

   - adjusting the gain (G) of the amplifier means (2) after processing each sample so that:

   * if a sample (E2) is outside the maximum range [-SMAX, +SMAX] the gain (G) is reduced by a predetermined amount;

   * if a predetermined number of successive samples (F0, F1; F4, F5; F8, F9) lie in the minimum range [-SMIN, +SMIN] the gain (G) is increased by a predetermined amount.
2. Method according to claim 1 characterised in that each sample is compared with four thresholds: a maximum threshold and a minimum threshold and their respective opposite values.
3. Method according to claim 1 characterised in that the absolute value of each sample is determined and then compared with the maximum threshold and the minimum threshold.
4. Method according to any one of claims 1 to 3 characterised in that the predetermined number of successive samples in the minimum range [-SMIN, +SMIN] for which the gain is increased is equal to 2.
5. Automatic gain control device (3, 4, 5) for a digital receiver (1), especially a time division multiple access (TDMA) digital receiver, said receiver (1) comprising variable gain amplifier means (2) receiving an input signal (e(t)), implementing the method according to any one of the preceding claims, characterised in that it comprises means (5) for sampling the output signal (s(t)) of said amplifier means (2) at a predetermined sampling frequency (Fe), means (4) for comparing the absolute value of each sample obtained with a maximum threshold (SMAX) and a minimum threshold (SMIN) and for generating threshold detection information, and means (3) for adjusting the gain (G) of the amplifier means (2), said adjustment means (3) cooperating with the comparator means (4) to reduce said gain (G) by a predetermined amount if a sample (E2) is detected outside a maximum range [-SMAX, +SMAX] and to increase said gain (G) by a predetermined amount if a predetermined number of successive samples (F0, F1; F4, F5; F8, F9) lie in a minimum range [-SMIN, +SMIN].
6. Device (3, 4, 5) according to claim 5 characterised in that the predetermined amounts by wnich the gain is reduced and increased are substantially equal to 3 dB.
7. Device (3, 4, 5) according to claim 5 or claim 6 characterised in that the minimum threshold (SMIN) is substantially 10 dB below the maximum threshold (SMAX).
8. Device (3, 4, 5) according to any one of claims 5 to 7 characterised in that the maximum threshold (SMAX) is substantially 6 dB below the saturation level of the baseband part of the receiver (1).
9. Device (3, 4, 5) according to any one of claims 5 to 8 characterised in that the sampling frequency (Fe) and the intermediate frequency (angular frequency ω_o) are chosen so that the ratio ω₀/Fe is an odd multiple of π/2 so that the variation in phase due to the carrier between two samples is an odd multiple of π/2, said sampling frequency (Fe) being further chosen to be very much higher than the symbol frequency of the input signal received so that between two samples the contribution of the modulation of the phase of the carrier by the flow of data is negligible in comparison with π/2.
10. Device (3, 4, 5) according to any one of claims 5 to 8 characterised in that the sampling frequency (Fe) and the intermediate frequency (angular frequency ω_o) are chosen so that the ratio ω₀/Fe is an odd multiple of π/2, said sampling frequency (Fe) being also chosen to be very much higher than the symbol frequency of the input signal received so that the envelope of said received signal varies slowly between two samples.

Images